Ubuntu
mountall package

boot failure because of missing filesystem event

Bug #533054 reported by Steve Magoun
12
This bug affects 1 person
Affects Status Importance Assigned to Milestone
mountall (Ubuntu)
Fix Released
High
Scott James Remnant (Canonical)
Karmic
Won't Fix
High
Unassigned
Lucid
Fix Released
High
Scott James Remnant (Canonical)

Bug Description

Binary package hint: mountall

While stress-testing Karmic, we ran into an infrequent boot failure that we traced to the 'filesystem' event not being emitted by mountall.

Keybuk describes the problem:

"mountall, in the failure case, has queued the "filesystem" event to be sent - but it's still queued only in memory and hasn't yet been put on the socket. Since mountall never waits for a reply, it's possible for mountall to exit without ever actually sending the event.

Adding:

  dbus_connection_flush (connection);

to the bottom of the main() function, just before the "return" should flush the sockets and ensure the event hits Upstart. If this fixes the problem, it's certainly something we'd want to SRU for karmic and fix in lucid!"

Revision history for this message
Steve Magoun (smagoun) wrote :

Note that mountall finished mounting the filesystems, so it should have emitted the event:
[ 6.299441] mountall: local finished
[ 6.299646] mountall: fhs mounted
[ 6.299816] mountall: mounted: local 2/2 remote 0/0 virtual 11/11 swap 1/1
[ 6.299916] event_new: Pending local-filesystems event
[ 6.300231] job_process_handler: Ignored event 1 (0) for process 470
[ 6.300412] mountall main process (470) exited normally
[ 6.300583] mountall goal changed from start to stop
[ 6.300753] mountall state changed from running to stopping

Scott James Remnant (Canonical) (canonical-scott)
Changed in mountall (Ubuntu):
status: New → Triaged
importance: Undecided → High
Changed in mountall (Ubuntu Karmic):
status: New → Triaged
importance: Undecided → High
assignee: nobody → Scott James Remnant (scott)
Changed in mountall (Ubuntu Lucid):
assignee: nobody → Scott James Remnant (scott)
Revision history for this message
Roman Yepishev (rye) wrote :

Can this bug be the reason why I get "Waiting for /home [SM]" in plymouth during boot that does not let to boot process to continue?
Once it has actually booted into gdm w/o /home mounted, once it stuck with "Waiting for /opt [SM]" - /home and /opt are separate partitions.
This is happening on a random basis and after several reboots I can get into a working environment, but I can't find any logs that might have helped to find out why it failed to mount /home.

Revision history for this message
Steve Magoun (smagoun) wrote :

Confirmed that the fix in the original bug report fixes the problem in our testing. With that fix, the filesystem event is reliably emitted.

Revision history for this message
Scott James Remnant (Canonical) (canonical-scott) wrote :

No, Roman; that's unrelated - your system is simply waiting for you to press either the S or M key. That's a bug in that it doesn't stop waiting when /opt does appear, but doesn't have anything to do when you press S or M.

Scott James Remnant (Canonical) (canonical-scott)
Changed in mountall (Ubuntu Lucid):
status: Triaged → Fix Committed
Changed in mountall (Ubuntu Karmic):
status: Triaged → Fix Committed
status: Fix Committed → Triaged
Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package mountall - 2.8

---------------
mountall (2.8) lucid; urgency=low

  * src/mountall.c:
    - Flush the D-Bus connection queue before exiting, otherwise we can
      end up leaving important events like "filesystem" in the memory queue
      without ever sending it on the socket. LP: #533054.
    - Move the message about disks being checked here from plymouth.
  * configure.ac:
    - Now that it's upstream, there's a proper ply-boot-client.pc file for
      the boot client library we want.
  * src/Makefile.am:
    - Which means we don't need to manually include -lplybootclient (the
      library changed names anyway)
  * debian/control:
    - Bump build-dependency
 -- Scott James Remnant <email address hidden> Fri, 12 Mar 2010 18:22:51 +0000

Changed in mountall (Ubuntu Lucid):
status: Fix Committed → Fix Released
Revision history for this message
Martin Pitt (pitti) wrote : Please test proposed package

Accepted mountall into karmic-proposed, the package will build now and be available in a few hours. Please test and give feedback here. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

Changed in mountall (Ubuntu Karmic):
status: Triaged → Fix Committed
tags: added: verification-needed
Revision history for this message
Scott James Remnant (Canonical) (canonical-scott) wrote :

Since no verification has been received, have given up with this SRU. Anyone affected can upgrade to Lucid soon enough.

Changed in mountall (Ubuntu Karmic):
status: Fix Committed → Won't Fix
assignee: Scott James Remnant (scott) → nobody
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.
Subscribing...

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.